As suggested by Vostro, I've clamped the Miller cap and now rail sticking is gone. Here are some scope traces that shows some nice curves under normal conditions as well as during clipping and hard clipping.
Frequency and amplitudes are shown on the graphs. Don't worry about the ripples visible on the traces, I think I have a parasitic problem into the scope...
For thse who doesn't like the square waves, note that these tests have been made at 40v p-p.
Even under hard clipping, never the amp has gone into oscillations, so that's good. I also join the exact schematic of the amp tested. The load was a 6r/300W resistor. The bias was set at 16mV between two output resistors.
BTW, I burned my speaker protection relay, under heavy clipping at 20KHz, the speaker-prot has triggered... and the contacts went into smoke (D'UH).
Enjoy!
There appears to be some rail sticking on the positive rail, worse at light clipping. There also seems to be an instability as the signal comes out of clipping. Something is ringing there. The output seems to be a a bit slew limited.
Do you have any series resistance between the function generator and the input (filter)? This will limit the SR of the square wave input. The bootstrap cap C9 has over 7V across it. When the output aproaches the positive rail, the bias on the MJE15033 CCS will be something like +62V on the emitter, +61.4V on base, but +70V on the collector. This might not be good for the MJE15033. What is the need for the bootstrap if you have a CCS to load the VAS?
Thank you CBS240
Giving us a helping hand.
I forgot the diodes in the output.... i have not imagined someone would be interested to clipping performance as in my imagination people will operate below clipping....but this is one more that i learn...people is interested in these details.
Living and learning...i have never imagined that... really...never!
Ahahahahha!.. i am really laughing about that
.... this is super precision...how dirty is the dirt...how messy is the mess....i do think we must run away from dirt or mess...not to study them...applied to clipping the same logics.....well....maybe i am wrong.,,,if before clipping i do think it is interesting to study.... but after clipping?... will people listen the clipping?... will be listening while clipping...will listen distortion by a pleasure?..... strange.
I have not watched how bad the clipping is....i have just inspected the performance till the amplifier clip... as people may reduce volume when perceiving clipping.... you know?.... what is the interest to know how an automobile engine is destroyed if you go in the red strip in the speedometer.. after the maximum RPM reached...engine will be without good lubrication..will be going to destruction.... i do not want to know...i just will avoid to destruct the engine and will operate in the blue safe region....well... this is my way of thinking.
I thought i was a little bit mad.... or little bit crazy..with one brain screw loosen.... i see that i am healthy... in comparison with people with concerns that way.
Merry Christmas boys...enjoy your life....huck the clipping!
Carlos
Giving us a helping hand.
I forgot the diodes in the output.... i have not imagined someone would be interested to clipping performance as in my imagination people will operate below clipping....but this is one more that i learn...people is interested in these details.
Living and learning...i have never imagined that... really...never!
Ahahahahha!.. i am really laughing about that
.... this is super precision...how dirty is the dirt...how messy is the mess....i do think we must run away from dirt or mess...not to study them...applied to clipping the same logics.....well....maybe i am wrong.,,,if before clipping i do think it is interesting to study.... but after clipping?... will people listen the clipping?... will be listening while clipping...will listen distortion by a pleasure?..... strange.I have not watched how bad the clipping is....i have just inspected the performance till the amplifier clip... as people may reduce volume when perceiving clipping.... you know?.... what is the interest to know how an automobile engine is destroyed if you go in the red strip in the speedometer.. after the maximum RPM reached...engine will be without good lubrication..will be going to destruction.... i do not want to know...i just will avoid to destruct the engine and will operate in the blue safe region....well... this is my way of thinking.
I thought i was a little bit mad.... or little bit crazy..with one brain screw loosen.... i see that i am healthy... in comparison with people with concerns that way.
Merry Christmas boys...enjoy your life....huck the clipping!
Carlos
Last edited:
Clipping behavior is a good test of unconditional stability. Eventually someone will crank it up and when they do, there should be no chance that an instability caused by that action might create failure of the output stage. I don't listen to clipping, but an amp should be able to drive a clipped signal for a period of time, or limit the input signal so as never to actually clip. This amp is close, a few minor tweeks ought to do it. This is where I take an educated guess as to where the issues are and do some trial & error.
Result of shorting R22
I tested the amp with R22 (VAS emitter resistor) shorted. The bad news is this did not eliminate the sticking to the negative rail. The good news is that it increased the output 6 volts rms before clipping begins. So the amp can now be driven harder without getting into clipping. Doesn't seem to be any downside to the change otherwise. I also looked closely at the waveform coming down from the positive peak. I didn't see any evidence of the glitch (ringing?) that Cannonica is observing. There are I believe several differences between my MK III and his including changes to the input stage emitter degeneration resistors (100R) and a higher global feedback factor. Perhaps the increased feedback is attenuating the glitch?
I'm interested in the comment by CBS240 about C9. I have to admit that I don't really understand it's function so I'm confused about it's relevance to the glitch problem. Perhaps someone would be kind enough to enlighten me.
I tested the amp with R22 (VAS emitter resistor) shorted. The bad news is this did not eliminate the sticking to the negative rail. The good news is that it increased the output 6 volts rms before clipping begins. So the amp can now be driven harder without getting into clipping. Doesn't seem to be any downside to the change otherwise. I also looked closely at the waveform coming down from the positive peak. I didn't see any evidence of the glitch (ringing?) that Cannonica is observing. There are I believe several differences between my MK III and his including changes to the input stage emitter degeneration resistors (100R) and a higher global feedback factor. Perhaps the increased feedback is attenuating the glitch?
I'm interested in the comment by CBS240 about C9. I have to admit that I don't really understand it's function so I'm confused about it's relevance to the glitch problem. Perhaps someone would be kind enough to enlighten me.
Clipping behavior is a good test of unconditional stability. Eventually someone will crank it up and when they do, there should be no chance that an instability caused by that action might create failure of the output stage. I don't listen to clipping, but an amp should be able to drive a clipped signal for a period of time, or limit the input signal so as never to actually clip. This amp is close, a few minor tweeks ought to do it. This is where I take an educated guess as to where the issues are and do some trial & error.
That's exactly my motivation...
The glitch on the positive side getting out of clip seems in the input stage. I'm trying to figure it out in the sim.
M.
I still advice to keep the diode in place, if you want: trial and error reducing the miller cap, that's what I would do.
The diode should not audiblily affect slew before clipping, if it affects slew after clipping, thats a mute point.
But interesting, I'll also do some research and listening tests.
I don't have access to my workbench till January, I can't think of how to get rid to the positive side glitch.
Hoping the more experienced members, will pitch in with a solution.
I would not remove the resistor (R22) by VAS, it has a function, side effect might be small reduction in max output Vpeak, but this difference is not audible.
4R7 in diagram is small, I use bigger resistor, but depends on VAS current. (so I guess I loose even more form my max output Vpeak)
@bonfis, what transistor are you using for VAS.
Kind Regards
The diode should not audiblily affect slew before clipping, if it affects slew after clipping, thats a mute point.
But interesting, I'll also do some research and listening tests.
I don't have access to my workbench till January, I can't think of how to get rid to the positive side glitch.
Hoping the more experienced members, will pitch in with a solution
.I would not remove the resistor (R22) by VAS, it has a function, side effect might be small reduction in max output Vpeak, but this difference is not audible.
4R7 in diagram is small, I use bigger resistor, but depends on VAS current. (so I guess I loose even more form my max output Vpeak)
@bonfis, what transistor are you using for VAS.
Kind Regards
Both Vostro... CCS sets the current mainly, but the bootstrapp
resistors are in series with the CCS feeding the VBE multiplier and the CCS colector to emitter junction.... we have some voltage drop there, but not so important than the CCS...mainly the CCS, but the bootstrap apply some signal back to the drivers.
I found sound better this way..with this combination.... also this is "marketing"..some guys believe in CCS, others in bootstrapp, this way i can make both sides happy.
Your flag Vostro?... what is your country?...you look latine by the Vostro (Italian meaning is "yours")... what is your native country despite you are in the other side of the fence.
I feel good vibrations from you man.... very good vibrations (sense from many others too..but yours are stronger and defined)...may i know something more from you...i see you want to help and not to show yourself...this is very good.... this is very positive.... want to help mankind not only to help yourself...i like that.
regards,
Carlos
resistors are in series with the CCS feeding the VBE multiplier and the CCS colector to emitter junction.... we have some voltage drop there, but not so important than the CCS...mainly the CCS, but the bootstrap apply some signal back to the drivers.
I found sound better this way..with this combination.... also this is "marketing"..some guys believe in CCS, others in bootstrapp, this way i can make both sides happy.
Your flag Vostro?... what is your country?...you look latine by the Vostro (Italian meaning is "yours")... what is your native country despite you are in the other side of the fence.
I feel good vibrations from you man.... very good vibrations (sense from many others too..but yours are stronger and defined)...may i know something more from you...i see you want to help and not to show yourself...this is very good.... this is very positive.... want to help mankind not only to help yourself...i like that.
regards,
Carlos
Last edited:
I still advice to keep the diode in place, if you want: trial and error reducing the miller cap, that's what I would do.
The diode should not audiblily affect slew before clipping, if it affects slew after clipping, thats a mute point.
But interesting, I'll also do some research and listening tests.
I don't have access to my workbench till January, I can't think of how to get rid to the positive side glitch.
Hoping the more experienced members, will pitch in with a solution
I would not remove the resistor (R22) by VAS, it has a function, side effect might be small reduction in max output Vpeak, but this difference is not audible.
4R7 in diagram is small, I use bigger resistor, but depends on VAS current. (so I guess I loose even more form my max output Vpeak)
@bonfis, what transistor are you using for VAS.
Kind Regards
I'm using the original MJE 15030. You mention a function for R22. maybe you could describe that for me. I noticed that some of Self's designs do not include this resistor.
@Mr Carlos, the flag is from Madagascar, but I'm from Scotish descent.
I came to this forum to learn about a passion of mine, electronics and audio, There are many many friendly and helpfull members on diyaudio, and I'm so pleased to be here.
I'm not an authority, but I build many circuits and try to scrutinise them in as much detail as I can, in the hope to learn as much as I can.
Knowledge is to be shared.
@bonfis, I've attached a pdf I put together for you, it's based on a example from Bob Cordells book, but I added as much detail as I could to make it an easy read.
Theres abit of maths, but very basic and only basic electronics knowledge needed, The formula can be taken as standard.
The circuits from D. Self arent always complete, since he deals with principles of the amplifier one stage at a time.
I wish he would post a diy circuit of his design for us.
Coming back to the amp, since we talking about degradation resistors, why not try altering them (make them higher), on VAS and Input.
lol But its a guess again
Output Iq (bias) will change and would need to be adjusted !!
Ive been to lazy to draw out the amplifier to simulate changes
Kind Regards,
I came to this forum to learn about a passion of mine, electronics and audio, There are many many friendly and helpfull members on diyaudio, and I'm so pleased to be here.
I'm not an authority, but I build many circuits and try to scrutinise them in as much detail as I can, in the hope to learn as much as I can.
Knowledge is to be shared.
@bonfis, I've attached a pdf I put together for you, it's based on a example from Bob Cordells book, but I added as much detail as I could to make it an easy read.
Theres abit of maths, but very basic and only basic electronics knowledge needed, The formula can be taken as standard.
The circuits from D. Self arent always complete, since he deals with principles of the amplifier one stage at a time.
I wish he would post a diy circuit of his design for us.
Coming back to the amp, since we talking about degradation resistors, why not try altering them (make them higher), on VAS and Input.
lol But its a guess again
Output Iq (bias) will change and would need to be adjusted !!
Ive been to lazy to draw out the amplifier to simulate changes
Kind Regards,
Last edited by a moderator:
I have the coil on my MKIII on the right way wrapped as Patrick and Hugh Dean suggested and that is a whole improvement.
I had the coil wrapped the wrong way, see #1462
Patrick, thanks for the right solution
Dx Blame MHIII Hx: https://www.youtube.com/watch?v=AHp82_KWuTQ
Regards,
Rudy
Merry Christmas
I had the coil wrapped the wrong way, see #1462
Patrick, thanks for the right solution
Dx Blame MHIII Hx: https://www.youtube.com/watch?v=AHp82_KWuTQ
Regards,
Rudy
Merry Christmas
The Dx Blame MKIII Hx amp sounds much better now did not expect this change gave this effect.
The MKIII Hx sound laid back read positive in compare to my Zap
Also I have other values resistors ordered for R23A and R23B
I am not know what value resistors to soldering, the 56R or the 100R
After that I skrew up the bias to 50mv
Regards,
Rudy
The MKIII Hx sound laid back read positive in compare to my Zap
Also I have other values resistors ordered for R23A and R23B
I am not know what value resistors to soldering, the 56R or the 100R
After that I skrew up the bias to 50mv
Regards,
Rudy
Last edited:
- Status
- Not open for further replies.